1. Field of the Invention
The present invention relates to an image processing technique for drawing an object having rough model data of multiple polygons and detailed model data of more polygons than the rough model data.
2. Description of the Related Art
Conventionally, in a video game, a virtual three-dimensional space is perspective-transformed onto a virtual screen from a predetermined viewpoint to draw an image of an object existing in the three-dimensional space. In order to display the progress of the game as a moving image, time in which one image must be drawn is limited to, for example, one frame period. For this reason, an amount of processing for drawing the image needs to be reduced. A Level Of Detail (LOD) technique is widely used to reduce the processing amount.
In the LOD technique, multiple model data, each having a different degree of detail for the same object (for example, a different number of polygons), is prepared. When a distance between a viewpoint and an object to be displayed is smaller than a predetermined threshold, an image of the object is drawn using detailed model data (for example, model data with a large number of polygons). When a distance between a viewpoint and an object to be displayed is larger than the predetermined threshold, an image of the object is drawn using rough model data (for example, model data with a small number of polygons).
According to the application of the LOD technique, an image of an object which is placed a short distance from the viewpoint can be displayed in detail using the detailed model data. An object which is placed a long distance from the viewpoint does not have to be displayed in detail, and an amount of processing for drawing the image can be limited by using the rough model data. For example, as shown in Unexamined Japanese Patent Publication No. 2003-115055, Unexamined Japanese Patent Publication No. 2003-228725, Unexamined Japanese Patent Publication No. 2001-250128, and Unexamined Japanese Patent Publication No. 2001-76180, a distance between a viewpoint and an object to be displayed is determined by a distance between a typical point, which is set at a central position of the object, and the viewpoint.
However, in the conventional LOD technique, when the distance between the typical point of the object and the viewpoint is larger than the threshold, an image of a part placed at a position much closer to the viewpoint is drawn using rough model data. When the distance between the typical point of the object and the viewpoint is smaller than the threshold, an image of a part placed at a position much farther from the viewpoint is drawn using detailed model data.
In the conventional LOD technique, when two or more such objects exist, there is a possibility that a part with a short distance from the viewpoint of one object will be drawn using rough model data, and that a part with a long distance from the viewpoint of the other object will be drawn using detailed model data. In this case, the position closer to the viewpoint is displayed roughly and the position farther from the viewpoint is displayed in detail. This gives an uncomfortable feeling to a player who watches the displayed image. Particularly, when multiple such objects exist close to each other, the player's discomfort increases.